Abstract
Near-infrared (NIR) Raman spectroscopy has been investigated as a tool to differentiate nasopharyngeal cancer (NPC) from normal nasopharyngeal tissue in an ex-vivo setting. Recently, we have miniaturized the fiber-optic Raman probe to investigate its utility in real time in-vivo surveillance of NPC patients. A posterior probability model using partial linear square (PLS) mathematical technique was constructed to verify the sensitivity and specificity of Raman spectroscopy in diagnosing NPC from post-irradiated and normal tissue using a diagnostic algorithm from three significant latent variables. NIR-Raman signals of 135 sites were measured from 79 patients with either newly diagnosed NPC (N = 12), post irradiated nasopharynx (N = 37) and normal nasopharynx (N = 30). The mean Raman spectra peaks identified differences at several Raman peaks at 853 cm-1, 940 cm-1, 1078 cm-1, 1335 cm-1, 1554 cm-1, 2885 cm-1 and 2940 cm-1 in the three different nasopharyngeal conditions. The sensitivity and specificity of distinguishing Raman signatures among normal nasopharynx versus NPC and post-irradiated nasopharynx versus NPC were 91% and 95%; and 77% and 96% respectively. Real time near-infrared Raman spectroscopy has a high specificity in distinguishing malignant from normal nasopharyngeal tissue in vivo, and may be investigated as a novel non-invasive surveillance tool in patients with nasopharyngeal cancer.
Original language | English |
---|---|
Pages (from-to) | 49443-49450 |
Number of pages | 8 |
Journal | Oncotarget |
Volume | 8 |
Issue number | 30 |
DOIs | |
Publication status | Published - 2017 |
Externally published | Yes |
Keywords
- Nasopharyngeal cancer
- PLS-DA model
- Raman spectroscopy
- Real time imaging
- Surveillance